CN1051099A - The digital speech coder that has optimized signal energy parameters - Google Patents
The digital speech coder that has optimized signal energy parameters Download PDFInfo
- Publication number
- CN1051099A CN1051099A CN90108421A CN90108421A CN1051099A CN 1051099 A CN1051099 A CN 1051099A CN 90108421 A CN90108421 A CN 90108421A CN 90108421 A CN90108421 A CN 90108421A CN 1051099 A CN1051099 A CN 1051099A
- Authority
- CN
- China
- Prior art keywords
- information
- value
- relevant
- parameter
- composition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000005540 biological transmission Effects 0.000 claims abstract description 41
- 238000012937 correction Methods 0.000 claims abstract description 11
- 239000000203 mixture Substances 0.000 claims description 33
- 239000013598 vector Substances 0.000 claims description 26
- 238000000034 method Methods 0.000 claims description 17
- 238000013139 quantization Methods 0.000 claims description 5
- -1 this Substances 0.000 claims description 2
- 230000005284 excitation Effects 0.000 abstract description 43
- 230000008878 coupling Effects 0.000 abstract 1
- 238000010168 coupling process Methods 0.000 abstract 1
- 238000005859 coupling reaction Methods 0.000 abstract 1
- 238000005086 pumping Methods 0.000 description 19
- 230000006870 function Effects 0.000 description 6
- 238000011084 recovery Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 239000004615 ingredient Substances 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000000295 complement effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000001105 regulatory effect Effects 0.000 description 2
- 238000001228 spectrum Methods 0.000 description 2
- 230000001944 accentuation Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 210000000697 sensory organ Anatomy 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L13/00—Speech synthesis; Text to speech systems
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
- G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques
- G10L19/08—Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters
- G10L19/083—Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters the excitation function being an excitation gain
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
- G10L19/04—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using predictive techniques
- G10L19/08—Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters
- G10L19/12—Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters the excitation function being a code excitation, e.g. in code excited linear prediction [CELP] vocoders
- G10L19/125—Pitch excitation, e.g. pitch synchronous innovation CELP [PSI-CELP]
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
- G10L2019/0001—Codebooks
- G10L2019/0004—Design or structure of the codebook
- G10L2019/0005—Multi-stage vector quantisation
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
- G10L2019/0001—Codebooks
- G10L2019/0011—Long term prediction filters, i.e. pitch estimation
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
- Analogue/Digital Conversion (AREA)
- Digital Transmission Methods That Use Modulated Carrier Waves (AREA)
Abstract
A kind of speech coder and code translator implementation, its pitch excitation and code table driving source energy are by parametric representation, and the desired transmission capacity of the transmission of these parameters can be reduced to minimum.These parameters are: energy value when long, one acts on when long energy value with correction factor and (or a plurality of) scale factor of coupling short-time energy, the relative size of the clear driving source of this (a bit) factor table and this short-time energy value.
Description
This invention relates generally to speech coder.Especially adopt the language of Gain Adjustable to express first digital speech coder.
Speech coder is well-known technology.Some speech coder is converted into digital expression with the artificial voice sample value, and then by adopting linear predictive coding to represent the spectrum information of language.Other some speech coders are then by providing the pumping signal relevant with original voice signal to improve common linear forecast coding technology.
United States Patent (USP) № .4,817,157 have described a kind of digital speech coder that improves vectorial driving source that has, the code table that a code table excitation vector is wherein arranged, can select the code table pumping signal that embodies the information that obtains most to its estimation, utilize it can provide and the immediate recovery voice signal of original signal.In such system, be used for producing the language information of restoring in conjunction with constituting composite signal thereby produce pitch ten excitation informations and code table excitation Xin Xi And and both.
Before these signal combination, each road signal all is provided with the energy ingredient that a gain factor is used to control each signal, thereby makes that the energy of each appropriate section is complementary in the energy ingredient of each signal and the original voice signal.This speech coder has also just been determined suitable gain factor when having determined excitation of suitable pitch and code table excitation information, can restore original language information thereby the coded message that includes all these elements offers code translator.Generally speaking, the speech coder of prior art is to provide these gain factor information with the form of disperseing to code translator.Finishing of this process is by above-mentioned packets of information transmission is realized that perhaps adopt other form (such as vector quantization) that they are combined, but this is the convenience in order to transmit, they still are separate separately compositions.
Previous speech coding technology has also been left over quite a lot of place that haves much room for improvement.The transmission mode of the gain factor that the front is mentioned may require transmission medium to possess sizable capacity to be used for error protection (otherwise error of transmission will be lost gain information, so also just causes the serious distortion of language information of recovery, is difficult to accept).
Therefore, be necessary to seek a kind of speech coding scheme, both reduced requirement, strengthened error protection simultaneously again gain factor information to transmission medium.
Here the speech coding scheme of Ti Chuing has solved above-mentioned problem basically.This encoding scheme will cause the generation of gain information, and this comprises that first is represented the first relevant yield value of composition and represents the second relevant yield value of composition with second of this language sample value with the language sample value.According to this scheme, these yield values will provide first parameter and one second parameter relevant with the whole energy of this sample value then through handling, this second parameter is based on first yield value and second yield value, or the value of one of them and the relative size of the whole energy values of this sample value at least.Information about this first and second parameter transfers to code translator then.
In one embodiment of the invention, gain information can comprise at least one with the 3rd relevant yield value of this sample value the 3rd composition gain, the processing of these yield values will produce one the 3rd parameter, this parameter determined at least in part first, second with the 3rd yield value in different sizes to the contribution of integral energy value.
In another embodiment of this invention: first, second parameter (if the words that the 3rd parameter exists be also included within) is passed through vector quantization so that a sign indicating number to be provided.This yard comprised the information that will be sent to code translator.
This invention be on the other hand: the gain information that scrambler produces comprises: first value that energy value is relevant when long with voice signal is (for example, the energy value that a plurality of sample value had, or the energy value that had of individual predetermined language information frame), with with signal short-time energy value (for example, signal sample or formation pre-determine the subframe of the part of frame) the second relevant value, this second value comprises a correction factor, it can be used for first value, to pass through specifying sample value or subframe to realize the adjusting that is worth first.What the transmission of first value from the scrambler to the code translator adopted is first rate, and second speed is adopted in the transmission of second value, and wherein second speed is more frequent than first rate.Like this, important information (energy value during length), the frequency of its transmission is low thereby its transmission can be adopted stronger safeguard measure and can not produce considerable influence to the transmission medium capacity.And the transmission of more inessential information (short-time energy value) is more frequent, but because their importance in signal recovers is smaller comparatively speaking, so also weak to the requirement of safeguard measure, also can reduce to minimum to the requirement of transmission medium capacity like this.
In another embodiment of this invention, the volume of language, decoding are installed in the middle of the wireless device.
Fig. 1 is a block diagram, has described corresponding to driving source structure of the present invention.
Fig. 2 also is a block diagram, has described the structure of the radio device relevant with this invention.
On March 28th, 1989 is be entitled as " digital speech coder with improved vectorial driving source " of Ira Gerson name issue, U.S. Patent number 4,817, in 157, described the digital speech coder that adopts vectorial driving source in detail, its vectorial driving source comprises a code table boot code vector code table.
This characteristic feature of an invention is to have adopted suitable digital signal processor (DSP) in speech coding (decoding) device, such as the DSP56000 family device of motorola inc.The calculation function of these DSP devices has been represented to come out as the equivalent electrical circuit block scheme in Fig. 1.
A pitch excitation filter state (102) provides a pitch pumping signal, and it comprises an instant pitch excitation vector.Multiplier (106) receives this pitch excitation vector ， And it is multiplied by the scaling factor of GAIN1.When correct realization, the energy that the weighting pitch excitation vector that is obtained is had will be corresponding with the energy of pitch information in the original speech information.Certainly, the energy of pitch information is different with initial sample value; The language sample value that will cause finally being restored than big-difference on the energy has the distortion of certain degree.
First code table (103) comprises a series of basis vectors, and their linearity is in conjunction with constituting a series of corresponding pumping signals.The function of scrambler is selected one in general exactly from these code table driving sources can characterize corresponding composition representative in the original language information.Code translator just utilizes scrambler institute to recover voice signal by select code table driving source.(certainly, for handled sample value, pitch pumping signal and code table are selected to define by corresponding composition to distinguish), corresponding pitch excitation information are to accept the code table excitation information and then be multiplied by the weighting factor of GAIN2 by multiplier (107).Being provided with of GAIN2 mainly is in order to regulate the energy of code table pumping signal, make its with corresponding language informational content in the actual energy of original signal consistent.
If necessary, the concrete application of this method also can utilize additional code table (104), and it contains additional pumping signal.The output of these additional code tables will need to regulate by a suitable multiplier (108), and suitable weighting factor (as GAIN3) can be realized purpose as hereinbefore.
After suitably selection and weighting were regulated, pitch excitation and code table excitation information addition (109) offered the LPC wave filter then and produce final voice signal.In scrambler, this final signal will compare with original signal, and other code table composition also need repeat this process, can identify a driving source like this, and final signal and original signal that it provided are the most approaching.This pitch and code table information are sent to code translator with Bei Bian Ma And by the transmission medium of selecting so.In decoder end, this final signal also will further be handled, and is the form that can listen with digitized information conversion, thereby finishes the recovery of voice signal.
For the description of characteristics of the present invention, we are first from the explanation decode procedure, and then illustrate from the angle of scrambler.
The function of gain control (101) provides the information (under some situation, also comprising the information of GAIN3) of GAIN1 and GAIN2.The actual energy of excitation of the pitch of this gain information and recovery and code table pumping signal by scrambler provide long the time energy value and the gain vector revised in short-term by the energy value when growing that scrambler provides etc. be closely related.
By pitch excitation filter state (102) and code table (103 and 104) (that is: early stage composition
*) the pitch pumping signal that is provided respectively and the energy of code table pumping signal can be determined by gain control (101) easily.Generally speaking, no matter the energy of these signals is the form of sharing between they two (or three) or provide with their form of total value, all can not correctly reflect the energy in the original signal.Therefore, be necessary to obtain this energy information so that determine required energy correction value.This energy correction is by regulating GAIN1 and GAIN2(if necessary, also comprising GAIN3) realize.Correction is that carry out each subframe on the basis with the subframe.
Carry out the pitch pumping signal and code table pumping signal energy calculation process has special advantage at code translator.Especially, the deviation of the pitch pumping signal energy that error of transmission in front caused can here be compensated, and this compensation is to realize by the accurate Calculation to the pitch excitation energy in the code translator.
For convenience of description, we suppose that an original language sample value (or its part) at least is digitized, and And and its final numerical information are to be divided into 3 Frames and subframe, and all these are consistent with well-known prior art.In the following description, we suppose that each frame contains 4 subframes.
For such structure, energy value occupies a subframe when long, and the corrected value in short-term that constitutes a correction factor is corresponding to a subframe.Being included in a residual amount of energy (EE) in the special subframe can determine by following formula generally speaking:
EE= (Eq(o))/(( FILTER POWER GAIN ) ( N - SOBS ))
Wherein:
E
q(o)=signal energy when growing corresponding to the digitizing of entire frame; FILTER POWER GAIN can obtain by calculating by the data of LPC wave filter, as is generally known it is corresponding to because energy increment that wave filter brought, N_SOBS represents contained number of sub frames in every frame.
Can calculate GAIN1 by following formula:
Wherein: α=primary vector parameter.
β=secondary vector parameter.
E
x(o)=unweighted pitch energy information.
Details about α and β also will be spoken of in the description of back encoding function.E
x(o) be the energy of the signal exported by pitch excitation filter state (102).Thereby, E
x(o) be exactly the energy that does not pass through the GAIN1 weighting at the pitch excitation vector of multiplier (106) front end.E
x(o) be on the denominator of A, the energy specification of the pitch excitation vector of weighting in the future turns to 1, and the molecule among the expression formula A has then comprised the energy of needs to the pitch excitation vector.In above-mentioned molecule, to mate the short-time energy in this pumping signal, β shows that then pitch excitation vector energy accounts for the ratio of this synthetic pumping signal energy to EE item (signal energy is to the estimated value of subframe residual amount of energy when long) by the α weighting.The square root of getting expression formula at last obtains this gain.
In like manner, GAIN2 can calculate by following formula:
α and β are ditto described, E
x(1) contain unweighted code table excitation information, this information is corresponding to the actual signal energy by first code table (111) output.
In case GAIN1 and GAIN2 determine as stated above, pitch excitation and code table excitation information be weighted appropriately so, thereby both values are complementary, and synthetic result is exported by totalizer (109), and suitable release signal composition is provided.In code translator, adopted one or more additional incentive code tables (104), additional weighting factor (as GAIN3) can be determined by same mode.
Characteristics of scrambler among the present invention will be described below:
As previously mentioned, for a whole frame of digital language sample value, can calculate a quantized signal energy value E
q(o).This value is sent to code translator by scrambler every now and then, to offer the information of code translator necessity.This information there is no need to follow each sub-frame information to transmit, because this frequency that information transmits when long is little, so this information can obtain comparatively safe protection by measures such as Error Corrections of Coding.Although this needs bigger transmission capacity, because the frequency of this information transmission is low, so little to the influence of whole transmission capacity.
And for example the front is described, and energy information needs during each subframe to do that corresponding to regulate the energy that makes with this frame the most approaching during corresponding to a frame long.This adjusting can be considered in short-term, and correction parameter α is the function of one of variable.
It equally also is the function of the energy ingredient of pitch excitation that scrambler is produced and code table excitation information signal as parameter that scrambler produces parameter alpha and β.α comprises a weighting factor, and energy information is through addend during according to this factor length, and pitch excitation information energy and code table 1 excitation and code table 2 are activated at addition in this subframe then.Parameter beta comprises a ratio, and this ratio is the pitch excitation information energy and the pitch excitation information of this subframe, code table 1 excitation, code table 2 excitation threes and ratio.Similarly, suppose that second code table exists, one the 3rd parameter π can express the first code table energy with pitch excitation information, code table 1 excitation, code table 2 excitation threes and ratio.
In above-mentioned processing mode, first parameter alpha is relevant with whole energy values of this signal sample, and second (if the words that have comprise the 3rd), parameter beta was relevant with the ratio of whole energy with one of pumping signal at least.Therefore, parameter alpha exists in a way relevant between β and the π three.This interrelated performance and coding and decoding efficient improved.
As one of characteristics, this Bian Ma Qi And is not with α, and three parameter values of β and π directly send code translator to.But, send code translator to but convert cognizance code to these three parameters process vector quantizations.Here because the code vector that scrambler spreads out of can not be equal to original vector fully, so may introduce error.For minimum is reduced in the influence that makes this error, scrambler to the vector sign indicating number that might obtain all calculate an ERROR value, select a vector that produces least error yard then.For being calculated as follows of each vector sign indicating number (it can produce corresponding α and β value, supposes it is solid size table scrambler here, is convenient to illustrate) this ERROR value:
naβ?+?λa(1-β)
Wherein:
In the middle of the superincumbent equation, Ev has represented the subframe energy of an ideal signal.Therefore, selected parameter is approaching more with initial parameter, and error is then more little.E
Pc(0) correlativity of the pitch information excitation of expression ideal signal and weighting.E
Pc(1) correlativity between the excitation of the code table of expression ideal signal and weighting.E
Cc(0,1) represents the excitation of weighting pitch information and the correlativity of weighting code table.Remaining, E
Cc(0,0) represents the energy of weighting pitch excitation, E
Cc(1,1) represents the energy of weighting code table excitation.(the weighting excitation is meant by the pumping signal after the sense organ accentuation filter processing of knowing).
After the vector sign indicating number with minimum ERROR value is differentiated out, this vector sign indicating number then is sent to code translator, after code translator receives this code vector, available it go to consult vectorial code data storehouse, thereby can recover α, β and π (if existence), as previously mentioned, these parameters will be used to calculate GAIN1, if GAIN2 and GAIN3(use it).
Adopt this scheme, can obtain several tangible benefits.For example: energy value when long thereby can guarantee that from the angle of energy information the language information of recovering is normal substantially because the safeguard measure in the transmission strengthens, and both has been correction factor information dropout or make mistakes that also relation is little in short-term.In addition, high-octane calculating of code translator middle pitch and compensation have reduced the error propagation of pitch excitation significantly.
The interrelated compression significantly that can allow information between parameter alpha, the initial gain information that β and π provided, thus make the required transmission capacity of this partial information transmission reduce to minimum.From effect, this programme has improved the Yu Sheng And that recovers and has reduced the transmission capacity requirement.
In Fig. 2, embody a cover radio device of the present invention and comprise an antenna (202) that receives language coded signal (201), the signal that a RF unit (203) processing receives is to recover the speech coding signal.This information offers the controlled variable that parameter code translator (204) produces each subsequent process thus.Driving source (100) utilizes the parameter generating pumping signal that offers it as previously mentioned.Pumping signal by driving source (100) output offers LPC wave filter (206), produces and the corresponding to synthetic video signal of coded message thus.This synthetic video signal recovers the quality of language then with enhancing by filtering (208) behind filtering (207) behind the pitch and the frequency spectrum.If desired, can also comprise that a postemphasis wave filter (209) further improves voice signal.This voice signal is exported the voice signal that can hear through handling then by speech convertor (212) at Audio Processing Unit (211).
The principal feature of the present invention in claims is as follows:
The information transmission scheme relevant with the gain information of single sample value, gain information wherein comprises:
With one first relevant yield value of first composition gain;
With at least one relevant second yield value of second composition gain;
Its feature shows as and has the following steps:
A) signal sample of the minimum limit of processing is to provide:
First parameter relevant with the total energy of these signal samples;
At least one second parameter relevant to small part and first and second yield values to the Relative Contribution of total energy value,
B) with the transmission information of first and second parameter correlations.
Gain information comprises one the 3rd yield value at least, and this value is relevant with the 3rd composition;
Processing procedure comprises that to provide one the 3rd parameter, this parameter to be based on different in first, second, third yield value Relative Contribution to the total energy value at least relevant;
The step of information transmission has comprised the transmission of Information that is associated with the 3rd composition.
Its treatment step comprises to major general's first parameter and second parameter information and carries out vector quantization so that a sign indicating number to be provided.
Its transmitting step comprises this sign indicating number of transmission.
Its also following transmitting step, i.e. relevant with a plurality of signal samples energy value information when long of transmission every now and then.
Its first parameter comprises a relevant correction factor of energy value information when long.
The feature of its transmitting step is to also have the following step:
B1). transmission and the relevant information of this first value every now and then;
B2). the transmission of Information relevant with second value is more frequent than the transmission of first value information.
The restored method of the information that gain information a kind of and each composition of signal is relevant, its feature shows as:
A) to receive first parameter relevant at least with the energy of at least a composition of this signal;
B) receive the composition definition information of this at least one composition;
C) handle this composition definition information, with provide one early stage composition, this, composition had an energy value in early stage;
D) if necessary, utilize the energy value of first parameter composition in this to revise at least, with the signal content that obtains restoring in early stage.
Claims (8)
1, the information transmission scheme relevant with the gain information of single sample value, gain information wherein comprises:
With one first relevant yield value of first composition gain;
With at least one relevant second yield value of second composition gain;
Its feature shows as and has the following steps:
A) signal sample of the minimum limit of processing is to provide:
First parameter relevant with the total energy of these signal samples;
At least one second parameter relevant to small part and first and second yield values to the Relative Contribution of total energy value,
B) with the transmission information of first and second parameter correlations.
2, the method in the claim 1, wherein
Gain information comprises one the 3rd yield value at least, and this value is relevant with the 3rd composition;
Processing procedure comprises that to provide one the 3rd parameter, this parameter to be based on different in first, second, third yield value Relative Contribution to the total energy value at least relevant;
The step of information transmission has comprised the transmission of Information that is associated with the 3rd composition.
3, the method for claim 1, its treatment step comprise to major general's first parameter and second parameter information carries out vector quantization so that a sign indicating number to be provided.
4, the method for claim 3, its transmitting step comprise this sign indicating number of transmission.
5, the method for claim 1, its also following transmitting step, i.e. relevant with a plurality of signal samples energy value information when long of transmission every now and then.
6, the method in the claim 5, its first parameter comprise a relevant correction factor of energy value information when long.
7, the method for claim 1, the feature of its transmitting step are to also have the following step:
B1). transmission and the relevant information of this first value every now and then;
B2). the transmission of Information relevant with second value is more frequent than the transmission of first value information.
8, the restored method of the information that gain information a kind of and each composition of signal is relevant, its feature shows as:
A) to receive first parameter relevant at least with the energy of at least a composition of this signal;
B) receive the composition definition information of this at least one composition;
C) handle this composition definition information, with provide one early stage composition, this, composition had an energy value in early stage;
D) if necessary, utilize the energy value of first parameter composition in this to revise at least, with the signal content that obtains restoring in early stage.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US42292789A | 1989-10-17 | 1989-10-17 | |
| US422,927 | 1989-10-17 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1051099A true CN1051099A (en) | 1991-05-01 |
| CN1097816C CN1097816C (en) | 2003-01-01 |
Family
ID=23676984
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN90108421A Expired - Lifetime CN1097816C (en) | 1989-10-17 | 1990-10-16 | Digital speech coder having optimized signal energy parameters |
Country Status (11)
| Country | Link |
|---|---|
| US (1) | US5490230A (en) |
| EP (1) | EP0570365A1 (en) |
| JP (1) | JPH05502517A (en) |
| KR (1) | KR950013371B1 (en) |
| CN (1) | CN1097816C (en) |
| AU (1) | AU652348B2 (en) |
| BR (1) | BR9007751A (en) |
| CA (1) | CA2065731C (en) |
| IL (1) | IL95753A (en) |
| NZ (1) | NZ235702A (en) |
| WO (1) | WO1991006943A2 (en) |
Families Citing this family (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| IT1241358B (en) * | 1990-12-20 | 1994-01-10 | Sip | VOICE SIGNAL CODING SYSTEM WITH NESTED SUBCODE |
| US5233660A (en) * | 1991-09-10 | 1993-08-03 | At&T Bell Laboratories | Method and apparatus for low-delay celp speech coding and decoding |
| US5692101A (en) * | 1995-11-20 | 1997-11-25 | Motorola, Inc. | Speech coding method and apparatus using mean squared error modifier for selected speech coder parameters using VSELP techniques |
| FI113571B (en) * | 1998-03-09 | 2004-05-14 | Nokia Corp | speech Coding |
| US6104992A (en) * | 1998-08-24 | 2000-08-15 | Conexant Systems, Inc. | Adaptive gain reduction to produce fixed codebook target signal |
| US7072832B1 (en) | 1998-08-24 | 2006-07-04 | Mindspeed Technologies, Inc. | System for speech encoding having an adaptive encoding arrangement |
| US6463407B2 (en) * | 1998-11-13 | 2002-10-08 | Qualcomm Inc. | Low bit-rate coding of unvoiced segments of speech |
| GB0005515D0 (en) * | 2000-03-08 | 2000-04-26 | Univ Glasgow | Improved vector quantization of images |
| US6754624B2 (en) * | 2001-02-13 | 2004-06-22 | Qualcomm, Inc. | Codebook re-ordering to reduce undesired packet generation |
| US7162415B2 (en) * | 2001-11-06 | 2007-01-09 | The Regents Of The University Of California | Ultra-narrow bandwidth voice coding |
| US7337110B2 (en) * | 2002-08-26 | 2008-02-26 | Motorola, Inc. | Structured VSELP codebook for low complexity search |
| CN101286320B (en) * | 2006-12-26 | 2013-04-17 | 华为技术有限公司 | Method for gain quantization system for improving speech packet loss repairing quality |
| US8688437B2 (en) | 2006-12-26 | 2014-04-01 | Huawei Technologies Co., Ltd. | Packet loss concealment for speech coding |
| US20090094026A1 (en) * | 2007-10-03 | 2009-04-09 | Binshi Cao | Method of determining an estimated frame energy of a communication |
| RU2586841C2 (en) * | 2009-10-20 | 2016-06-10 | Фраунхофер-Гезелльшафт цур Фёрдерунг дер ангевандтен Форшунг Е.Ф. | Multimode audio encoder and celp coding adapted thereto |
| US8862465B2 (en) | 2010-09-17 | 2014-10-14 | Qualcomm Incorporated | Determining pitch cycle energy and scaling an excitation signal |
| US20150173473A1 (en) * | 2013-12-24 | 2015-06-25 | Katherine Messervy Jenkins | Convertible Activity Mat |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| NL8500843A (en) * | 1985-03-22 | 1986-10-16 | Koninkl Philips Electronics Nv | MULTIPULS EXCITATION LINEAR-PREDICTIVE VOICE CODER. |
| US4969192A (en) * | 1987-04-06 | 1990-11-06 | Voicecraft, Inc. | Vector adaptive predictive coder for speech and audio |
| US4868867A (en) * | 1987-04-06 | 1989-09-19 | Voicecraft Inc. | Vector excitation speech or audio coder for transmission or storage |
| US4910781A (en) * | 1987-06-26 | 1990-03-20 | At&T Bell Laboratories | Code excited linear predictive vocoder using virtual searching |
| US4899385A (en) * | 1987-06-26 | 1990-02-06 | American Telephone And Telegraph Company | Code excited linear predictive vocoder |
| US4817157A (en) * | 1988-01-07 | 1989-03-28 | Motorola, Inc. | Digital speech coder having improved vector excitation source |
| EP0331857B1 (en) * | 1988-03-08 | 1992-05-20 | International Business Machines Corporation | Improved low bit rate voice coding method and system |
-
1990
- 1990-09-24 IL IL9575390A patent/IL95753A/en not_active IP Right Cessation
- 1990-10-09 EP EP90915602A patent/EP0570365A1/en not_active Withdrawn
- 1990-10-09 CA CA002065731A patent/CA2065731C/en not_active Expired - Lifetime
- 1990-10-09 BR BR909007751A patent/BR9007751A/en not_active IP Right Cessation
- 1990-10-09 WO PCT/US1990/005693 patent/WO1991006943A2/en not_active Application Discontinuation
- 1990-10-09 JP JP2514552A patent/JPH05502517A/en active Pending
- 1990-10-09 AU AU66031/90A patent/AU652348B2/en not_active Expired
- 1990-10-09 KR KR1019920700861A patent/KR950013371B1/en not_active IP Right Cessation
- 1990-10-16 CN CN90108421A patent/CN1097816C/en not_active Expired - Lifetime
- 1990-10-16 NZ NZ235702A patent/NZ235702A/en unknown
-
1994
- 1994-12-22 US US08/361,474 patent/US5490230A/en not_active Expired - Lifetime
Also Published As
| Publication number | Publication date |
|---|---|
| US5490230A (en) | 1996-02-06 |
| AU652348B2 (en) | 1994-08-25 |
| IL95753A (en) | 1994-11-11 |
| EP0570365A1 (en) | 1993-11-24 |
| WO1991006943A3 (en) | 1992-08-20 |
| CA2065731C (en) | 1995-06-20 |
| NZ235702A (en) | 1992-12-23 |
| KR920704266A (en) | 1992-12-19 |
| KR950013371B1 (en) | 1995-11-02 |
| EP0570365A4 (en) | 1993-04-02 |
| CN1097816C (en) | 2003-01-01 |
| IL95753A0 (en) | 1991-06-30 |
| JPH05502517A (en) | 1993-04-28 |
| AU6603190A (en) | 1991-05-31 |
| WO1991006943A2 (en) | 1991-05-16 |
| CA2065731A1 (en) | 1991-04-18 |
| BR9007751A (en) | 1992-07-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1135721C (en) | Audio signal coding method and apparatus | |
| US6826526B1 (en) | Audio signal coding method, decoding method, audio signal coding apparatus, and decoding apparatus where first vector quantization is performed on a signal and second vector quantization is performed on an error component resulting from the first vector quantization | |
| CN1051099A (en) | The digital speech coder that has optimized signal energy parameters | |
| DE60120766T2 (en) | INDICATING IMPULSE POSITIONS AND SIGNATURES IN ALGEBRAIC CODE BOOKS FOR THE CODING OF BROADBAND SIGNALS | |
| CN1123866C (en) | Dual subframe quantization of spectral magnitudes | |
| CN1121683C (en) | Speech coding | |
| CN101996636B (en) | Sub-band voice codec with multi-stage codebooks and redundant coding | |
| CN1154086C (en) | CELP transcoding | |
| CN1051392C (en) | Vector quantizer method and apparatus | |
| EP1914724B1 (en) | Dual-transform coding of audio signals | |
| CN1288621C (en) | Error concealment in relation to decoding of encoded acoustic signals | |
| CN1225723C (en) | Noise suppression | |
| CN1334952A (en) | Coded enhancement feature for improved performance in coding communication signals | |
| CN1092420C (en) | Audio data decoding apparatus and audio data coding/decoding system | |
| EP2037451A1 (en) | Method for improving the coding efficiency of an audio signal | |
| CN1470050A (en) | Perceptually improved enhancement of encoded ocoustic signals | |
| CN1193344C (en) | Speech decoder and method for decoding speech | |
| CN1661924A (en) | Audio codec system and audio signal encoding method using the same | |
| CN1192357C (en) | Adaptive criterion for speech coding | |
| CN1173690A (en) | Method and apparatus fro judging voiced/unvoiced sound and method for encoding the speech | |
| CN114550732A (en) | Coding and decoding method and related device for high-frequency audio signal | |
| CN1124588C (en) | Signal coding method and apparatus | |
| CN101562015A (en) | Audio-frequency processing method and device | |
| JP4359949B2 (en) | Signal encoding apparatus and method, and signal decoding apparatus and method | |
| CN1132157C (en) | Linear predictive analysis-by-synthesis encoding method and encoder |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CX01 | Expiry of patent term |
Expiration termination date: 20101016 Granted publication date: 20030101 |